Research On Key Issues In Digital Design Platform For Parallel Robots

High speed, high precision and intelligence are common development trends for parallel robots. However a number of problems are not well solved in the design and development of parallel robots, i.e. long development cycle and high development cost. This research is sponsored by two National projects on integrated robot design.Considering the demands of advanced engineering analysis and fluent process management in the forward design process of robots, a novel software development framework which combines several CAD/CAE integration technologies is first proposed. Then a digital R&D platform for parallel robots is developed employing modern software technologies. The platform is able to cover most important aspects of parallel robots design, including conceptual design, structural design, components selection, various performance prediction, and design verification, etc., and it also promotes better management of design process and relevant data. The following contributions have been made:For engineering design(i.e. CAESAM) and management and related digital design technologies(i.e. Sys ML) are summarized and analyzed. With an in-depth examination of the kernel concepts of parallel robots design, a novel framework for the digital design platform is proposed and major GUI layouts and GUI elements are conceived.The novel framework is developed utilizing advanced software technologies like.Net framework, plug-in framework, and dynamic language runtime, and hence provides flexible and powerful interfaces which support a better of integration between the main program and other software like Excel, Matlab, SolidWorks,SAMCEF etc. And most existing analysis and post-processing tools can be integrated.This research also proposes a new method of data exchange and management based on Modelica text and Excel macro files. Based on the method, a parameter driven framework for CAD-MBS-CAE is constructed, e.g. parametric models in Solid Works,Modelica and SAMCEF are constructed and mapped. Further a consistent management of “Model-Data-View” linkages among different design stages is realized.Several important supporting databases demanded in the integrated design process of parallel robots are developed. A library of dynamics models for typical parallel robots is built based on the Modelica language and corresponding solver software. The low precision issue of existing research which uses simplified Modelica models for parallel robots is solved for the first time.The research investigates systematically novel integration methods between existing robotic control algorithms, e.g. the combination of ADRC(Active Disturbance Rejection Control) and traditional control algorithms. The effects of ADRC, input shaping, PDFF-I/II on the precision of robot trajectory and the effects of vibration suppression are experimented. Some rules of thumbs for customized control algorithm integration are proposed. All the algorithms are coded into the robot control algorithm library.RoboDev, the digital design platform of robots, is developed based on above technologies. The validity of our integration methods and our supporting databases is proved by simulations of typical parallel robots.